Deep Learning as a Black Box, “OpenAI” and the Internet. On the media-political dimension of machine learning methods
Internet-based digital culture is increasingly determined by modern artificial intelligence technologies. Especially approaches of so-called Deep Learning, which are oriented towards the functioning of biological brains, are currently booming. As machine learning methods, they make it possible for a computer to learn without being explicitly programmed for this purpose. The research project analyses the media-political dimension of machine learning methods. Two aspects are at the centre of this: on the one hand, there is the question of the extent to which deep learning represents a black-box technology, i.e. it is to be regarded as a system whose internal operations and functions are closed to human understanding. On the other hand, industrial-scientific projects (e.g. the non-profit research company OpenAI) are to be examined, which, at least according to their own claim, strive for a democratisation of artificial intelligence. The project not only examines whether these claims are being fulfilled. Rather, it analyses the different medial conditions of machine learning and situates them theoretically, historically and politically.
Research focus
- Media history of machine learning
- Aesthetics and politics of popular and digital image media
- Media criticism
- Forms and processes of the documentary
Curriculum Vitae
- since 2017: Private lecturer at the Institute for Media Studies at the Ruhr University Bochum
- July 2016: Habilitation and Venia legendi in media studies at the Ruhr University Bochum
- 2013 – 2016 Head of the DFG sub-project “Digital Seriality: Serial Aesthetics and Practice of Digital Play” within the framework of popular seriality” at the John F. Kennedy Institute, Freie Universität Berlin, together with Shane Denson.
- Previous positions: e.g. Acting Head of the Department of Media Studies at the Centre for Interdisciplinary Media Studies (ZiM), Georg-August-Universität Göttingen, and Research Associate at the Institute for Media, Information and Cultural Studies (IMIK), University of Regensburg.
Publications
Monographs:
- Sudmann, A. (2017). Serial overbidding. On the televisual aesthetics and philosophy of exposed increases. Stuttgart: Metzler. (forthcoming, also habilitation thesis, July 2016).
- Sudmann, A. (2006). Taming the Shrew? On the Politics of Representation in Contemporary US Independent Film. Bielefeld: transcript.
- Sudmann, A. (2001). Dogma 95: Turning away from the constraint of the possible. Hanover: Offizin-Verlag.
Editorships:
- Sudmann, A. & Engemann, C. (2017). Machine Learning. Media, infrastructures and technologies of artificial intelligence. Bielefeld: transcript.
- Sudmann, A. & Denson, S. (2014). Digital Seriality. Journal for Computer Game Culture 8.1. special-issue. Online.
- Sudmann, A. & Strobel, R. (2009). Film transnational and transcultural. European and American Perspectives. Paderborn: Wilhelm Fink.
- Sudmann, A. & Stockmann, R. (2008). Computer Games as a Sociocultural Phenomenon. Games Without Frontiers – War Without Tears. Basingstoke; New York: Palgrave Macmillan.